06/18/01 1

Cryogenic Roadmap

U.S. Department of EnergySuperconductivity Program for Electric Systems

Editors: Thomas P. Sheahen, Ben McConnell

This document, originating within the national laboratories and their industrialcollaborators, presents a series of suggestions for research and development pertaining tocryogenics. Appendix A is a tabulation of the steering committee members whoparticipated in the development of this document. The editors have primarily used thematerials and discussions provided by the steering committee and the 1997 and 1998cryogenic workshop findings. However, there are added clarifications and suggestedoutcomes that were not presented in the workshop summaries.

Cryogenic systems providing 100-1000 watts of cooling at 65-80 K are required ifdevices utilizing high-temperature superconductors are to become a part of the nationalelectric power delivery and utilization system. These systems must have lower capitalcosts and operate more efficiently and reliably then current off the shelf cryogenicsystems. In addition, the physical size, maintainability, and operation must not constrainthe expected benefits of high temperature superconducting power equipment. Thisdocument addresses the various pathways for the development of cryogenic systems thatwill enable cryogenic systems to advance from the present state-of-the-art to systemsmeeting these desired characterisitics. Consequently, it is called a “roadmap”. Theroadmap provides goals and objectives along with the desired outcomes that may result ifthese goals and objectives are completed.

Introduction & Background

The new technology of High Temperature Superconductivity (HTS) offers thepromise of many applications, especially that of improving the efficiency of electricpower generation, transmission, distribution and use (in electric motors) by means ofcurtailing the i2R loss that inevitably occurs in conventional technology used today.Besides improving efficiency, HTS technology also offers the prospect of increasingcapacity per unit volume (higher power density) in comparison to conventional electricalequipment. This would allow more efficient use of utility rights-of-way and othereasements.

The U.S. Department of Energy (DOE), through its Superconductivity PartnershipInitiatives (SPI) program, is sponsoring teams of collaborators made up of nationallaboratories, industry and utilities who are building and testing prototype devices for usein the electric power system. The array of projects includes: transformer, motor, cable,fault-current limiter, and flywheel energy storage.

06/18/01 2

In most of these initial SPI projects, cryogenic support has not been a key focusfor the project; that is, cooling to operating temperature has been achieved viarefrigeration equipment sufficient to get the job done, usually disregarding questions ofthe cooling system efficiency, reliability or cost. (The ABB HTS transformer project isexceptional in that it is evaluating the refrigeration needs as an integral part of thesystem.) Everyone recognizes that the actual use of such devices in the real world ofelectric power will require cooling systems that are indeed an integral part of the system.For this reason, there will be further refinement of cryogenic systems accompanyingfuture programs, perhaps in another round of SPI projects someday. In anticipation ofthat, it is desirable at the present time to gain an understanding of what needs to beaccomplished in the area of cryogenics; hence the need for a roadmap.

There have been two recent workshops held by DOE to take the first stepstoward such a roadmap:

The first was held July 22, 1998 in Washington DC; Appendix B presents a briefsummary of its outcome. Attendees at that conference examined the present state of theart in cryogenics and outlined the performance-parameters of the equipment that will beneeded if HTS devices are to “come true” as part of the national electric power system.The full conference proceedings are entitled Cryogenics Needs of Future HTS ElectricalPower Equipment, and comprise a key document underlying this roadmap.

The second workshop took place on July 27, 1999. Here the emphasis was onfinding out what are the most important considerations of both researchers and users withregard to actually accomplishing the goal of providing cryogenics to meet the needs of thefuture electric power system, when it includes HTS devices. Appendix C presents a briefsummary of that conference.

The three evaluation criteria of cryogenic performance are: efficiency, reliabilityand cost. In any utility application, reliability is absolutely indispensable, and any devicethat even raises questions of reliability will be rejected out of hand. In general, higherreliability would imply higher costs, especially if reliability is achieved throughredundancy (dual systems). Weighing the trade-offs between these three criteriaconstitutes a key element in the design process of any cryogenic system. The finalconfiguration must optimize these criteria in a way that enables the HTS devices tocontribute profitably to the operation of the national electric system.

Present Status

Existing HTS devices need to be maintained at 25K to 77K. (With technologicaladvances in HTS wire, future devices may operate near 80K, which would greatly reducethe cost of cooling.) Cryogenic refrigeration systems can be categorized as two basictypes: closed loop systems that use a cryocooler to provide refrigeration, and open loopsystems that use once-through cryogens to provide refrigeration. A hybrid system usesboth of these.

For small-scale applications, present economics clearly favor vaporization of LN2

over mechanical refrigerators. Only when temperatures below the operating range of LN2

06/18/01 3

(T < 65 K) are required will small-scale mechanical systems be applied. Thus our interestin closed-loop systems is focused on intermediate and large sizes.

A. Closed-Loop Systems – Mechanical Refrigerators

The basic science of cryogenics systems pertinent to HTS applications isadequately presented elsewhere1,2,3. The practical engineering of mechanical devices ismuch more complicated. There are many different types of refrigerators, each of which isbest in a certain range.

Figure 1 – Refrigeration power requirements and applications.

Figure 1, taken from Radebaugh’s presentation1, displays using variously shadedregions of temperature and refrigeration power where one or another technology worksbest. Circles indicate where certain of today’s SPI devices fall on this graph; (these circlesmust not be confused with the long term devices that will meet the needs of the utilitysector). The problem of selecting the “best” refrigerators for the HTS devices that servemost electric power applications can be understood by scrutinizing the right-central areaof figure 1: at 77 K, the desired refrigeration power (100 – 1000 W) is slightly below theTurbo-Brayton Cycle range and slightly above the Gifford McMahon range. Pulse Tubesand Stirling Cycle refrigerators cover this range, but only a limited number of suchsystems have been produced, and their track record in industrial applications is limited.Pulse tube and Stirling cryocoolers developed for space applications have reached thisreliability (10 years continuous operation for a few Stirling coolers) but the costs are veryhigh at present. Such reliability is much better than Gifford-McMahon refrigerators.Accordingly, there is a need for improvements in technology that will cover the operatingrange of greatest interest to us.

The efficiency of mechanical cryogenic refrigerators has been improving steadilyover many years, but for the devices of greatest interest to HTS applications, theirefficiency is typically <20% of Carnot efficiency, defined by:

ηc = Tc / (Th – Tc).

For an ambient temperature Th = 300 K and an operating temperature Tc = 77 K,ηc = 0.345; so a typical efficiency is 20% of that, or 6.9%. Another way to represent

this is through the Specific Power ( = 1/η ), which is 14.5 in this case. (Specific Power isexpressed in units of watts per watt, i.e. the number of watts of input power required toremove one watt of heat at the cooling temperature.) This means that to remove one watt

06/18/01 4

of heat from the 77 K region, it requires 14.5 watts of input energy. This requirement isknown as the cryogenic penalty, because it is equivalent to a parasitic loss that diminishesthe savings obtained by using HTS wire and eliminating the i2R loss of the particularelectric device.

A variety of calculations have been done to model the behavior of HTS devicesoperating at various temperatures, including 30 K and 64 K. Obviously the Carnotefficiency is lower at those lower temperatures; e.g., at 30 K, ηc = 1/9. Hence the actualoperating efficiency is lower, the specific power of the refrigerator is higher, and thecryogenic penalty is more severe.

Calculations by Mulholland et al4 suggest that for a 1000 hp electric motoroperating at 30 K, the energy savings (i2R , etc.) are insufficient incentive to compensatefor the cryogenic penalty. However, Blaugher5 has calculated that a 5000 hp motor islarge enough to overcome the cryogenic penalty at 30 K, giving a net improvement inefficiency of 1.7%. Transformers and generators save a smaller percentage of the deviceenergy, and hence are even more sensitive to the cryogenic penalty, and to the operatingtemperature. Consequently, for most electric-power applications, it is an importantresearch goal to develop second-generation HTS conductors that carry high current attemperatures near 77 K.

Therefore, this roadmap confines its attention to refrigerators that will operateabove 60 K, with greatest emphasis given to 77 K cryogenic systems. Table 1, providedby American Superconductor Corp., is a tabulation of key parameters of existingcryogenic systems. It will be immediately recognized that the capital cost per watt isvery high, and the efficiency (expressed as a percent of Carnot efficiency) is quite low.Clearly there is plenty of room for progress.

Discussion at the first (1998) workshop (see Appendix B) touched in part on thecondition characterized by the phrase “If you build it, they will come.” Cryogenicmanufacturers have already expressed confidence that it is possible to build excellentrefrigerators that will meet the needs of these HTS devices, but without a substantialmarket, they have no incentive to do so. There is a “chicken & egg” effect going on,whereby the HTS market will always remain small until refrigerators are available, but

06/18/01 5

refrigerators will not become available until the market for HTS devices grows. It is not

Refrigeration Approximate Capital Cost Input Power/Temperature Working Fluid Capacity Refrigeration Cycle Expander Plant Cost per Watt Input Power Cooling Pwr % Carnot

(K) (Supplier Units) (W) $$$ (KW) (W/W)

Large Scale GM Systems Leyblod 120T 65 He 130 W 130 GM Recip $20,200 $155.38 6.5 50.0 7.23% Cryomech AL200 65 He 150 W 150 GM Recip $19,600 $130.67 5.5 36.7 9.86%

Helium Gas Systems PSI Model 1620 65 He 1200 W 1200 Claude Recip $400,000 $333.33 105 87.5 4.13%

Stirling Cycle Stirling Cryogenics LPC04 80 He 4200 Watts 4200 Stirling Recip $400,000 $95.24 45 10.7 25.67%

Large Scale Recondensing Systems PSI 80 ?? 11,500 Watts 11,500 Brayton Turbine $800,000 $69.57 167 14.5 18.94%

Liquid Air Plants Cosmodyne GF-1 80 N2 4 T/Day 8,400 Brayton Turbine $700,000 $83.33 372 44.3 6.21%

Cosmodyne Aspen 1000 80 N2 1000 nM3/Hr 64,969 Brayton Turbine $2,650,000 $40.79 1400 21.5 12.76%

GEECO EDLP-20TN 80 N2 20 T/Day 42,000 Brayton Turbine $1,790,000 $42.62 933 22.2 12.38%

GEECO EDLP-40TN 80 N2 40 T/Day 84,000 Brayton Turbine $2,750,000 $32.74 1773 21.1 13.03%

Table 1 – Present Commercial Cryogenic Systems

at all clear how this problem can be overcome. Certainly government-sponsored R&D isnot customarily carried beyond demonstrating one or two devices to establish feasibility;the expansion to a wider market is left to entrepreneurs in the private sector.

Open Loop Systems - Liquid Nitrogen Based

There is a large liquid nitrogen production and supply infrastructure in placethroughout the world with literally thousands of tons of available LN2. This can beutilized in HTS applications with very attractive economics. Such systems would besimple to install and operate with very few complex components. Temperatures in therange of 65K to 77K can be achieved with vacuum pumps.

There are many apparent advantages to an open loop system with LN2

evaporating in pool boiling. The capital costs of such systems are low, and operationalreliability is significantly higher since no major machinery is needed and it is a simplesystem. On the other hand, while LN2 costs are very competitive and the operatingeconomics look superior to small mechanical refrigeration systems, costs could be higherin some remote locations. Also, geometries of some systems may not favor forced circulation cooling over pool boilingand therefore may need special cryostat designs.

06/18/01 6

There are several areas for improvement in existing LN2 open loop coolingsystems, and a program that strives toward those improvements could help optimize theHTS refrigeration system

Desired Future Status – Goals, Objectives and Outcomes

A basic goal of any sponsored research is to work the sponsor out of a job. Byproviding seed money, the sponsor strives to nucleate a commercial developer. Like anygood R&D program, the quest for better cryogenic systems is intended to reach a statewhere manufacturers are competing with one another in a major marketplace, meetingcustomers’ requirements without benefit of any subsidy from a sponsor.

For many other applications of cryogenics, this state has already been reached.However, it appears (based on the workshops of 1998 and 1999) that for the case of HTSdevices operating within a utility system, the near-team market outlook is too small andtoo specialized to attract a fully independent effort by existing cryogenic manufacturersto meet the performance requirements associated with these devices. It is thereforeappropriate to discuss the kind of R&D program that could lead to HTS devices beingbought and sold on the same business footing as other commercial products.

Direction

The very fundamental question must be asked: What are we trying to do? Forcryogenics, the answer is “Remove heat.” In any electrical power device or system, theload current varies over time. In general, for those HTS systems in which thesuperconductor carries varying current, the cooling load that the cryogenics must handlealso varies over time. The cryogenic system must be sized to accommodate the highestload, not just the average load. That can become expensive in some instances, requiringlarge up-front capital outlays for a cooling capability that is fully used only infrequently.Multiple staged cryogenic systems offer one expensive solution to this problem. Avariable capacity cryogenic system is a much more attractive oprion.

For systems operating near 77 K, the choice of a liquid nitrogen bath is aconvenient way to meet both peak and average loads. (Keeping a LN2 storage tank filledis a way to deal only with the “average” load.) However, any system operating around64 – 70 K demands additional cooling equipment (perhaps a vacuum pump; perhaps anelectric cryocooler operating from an “ambient” of 77 K, etc.), and that must be sized tohandle the maximum load. Moreover, not all applications are able to use a LN2 bath, forgeometric and dielectric reasons.

Recognizing that most HTS devices must operate in a utility setting, which canrange from an urban power station to an isolated distribution substation in “desert”condition, the cryogenic system must conform to the parameters of performance placedupon other subsystems within the utility environment. These parameters include veryhigh reliability/availability, low maintenance, and automatic/remote operation. Withoutthese, any HTS device will simply not earn acceptance in the marketplace.

06/18/01 7

Large electric motors (5000 hp and up are candidates for HTS technology) arecommonly used in factories, and have slightly less stringent performance requirementsthan utilities. However, in many industrial applications the shorter payback periodforces the cost parameter to be more stringent than for utilities. Overall, the demandsupon cryogenic systems are very severe.

Goals for Major Parameters

It is easy to say that everyone wants cryogenic refrigerators that are efficient,cheap and reliable relatively small in size and acceptable to utility and industrial markets.The much more demanding task is to turn these adjectives into quantitative measures ofperformance. Toward that end, we propose the following major goals: an efficiency of30% relative to Carnot, a capital cost below $25 per watt of cooling, and anavailable operating time (or reliability) exceeding 99.8%. Table 2 below summarizesthe goals, objectives and desired outcomes. As discussed below, these goals are reasonablewhen compared to the present status of cryogenic systems and when weighed against thepotential outcomes are highly desirable.

The improvement of efficiency from today’s range, 20% of ηc, to 30% appears tobe a tractable goal, neither too ambitious nor too conservative. This objective is entirely atechnical one.

The distinction between Availability and Reliability deserves clarification.Reliability, in a strict sense, is a measure of the fraction of units operating after a givenperiod of time. The average lifetime of a refrigerator deals with the time until replacement.Availability is the more important parameter to those who have profitability tied to beingable to meet customer specifications. Utilities are generally much more concerned withavailability measures than reliability per se. The numerical goal for availability may seemstunning by laboratory and factory standards, but not so for utility systems. In fact, thenumber of 99.8% may be low for many applications in real utility systems.

For example, if a HTS cable brings power into certain blocks of a dense centralcity where there is no alternate delivery path, failure of the cryogenics implies a blackout.Virtually all utilities will insist upon both scheduled maintenance and redundancy for theforeseeable future, until HTS systems are proven beyond any doubt. The achievement ofthis goal is absolutely indispensable for utility applications.

When building a real cooling system, cost and reliability are strongly coupled. Thetrade-off of cost for availability that is associated with redundancy is familiar to anyutility engineer. For example, one report6 mentions a case where six standard nitrogenplants had an average of 18 outages per year. However, due to redundancy, this resultedin nitrogen delivery failure only once in 30 years. The history of the utility industryshows that redundancy is the commonest way to assure very high availability.

Redundancy helps to ensure availability, but any instances of multiple outages ofa system are unacceptable to the utility. The breakeven point between an affordabledevice (perhaps one risking failure) and a more expensive, more reliable device must be

06/18/01 8

considered – not only from the “maintenance” viewpoint, but also from the perspectiveof user confidence in the system. Risk carries a very high cost penalty.

Cryogenic manufacturers have worked hard over many years to improvereliability, by eliminating moving parts at the cold end and reducing mechanical wearthrough use of flexure bearings and gas bearings. In one representative (pulse-tube) case7,the top three life-limiting phenomena were helium leakage, gas contamination, and powerelectronics failure. The first two are controllable through proper design (and perhaps aregularly scheduled helium clean- and fill-up). That leaves the power electronics, the“standard” against which most people judge reliability. To meet utility requirements,manufacturers customarily start with a very reliable design and then try to decrease thecost.

Today’s cryogenic technology is very expensive. Table 1 suggests a ballparkfigure of $100/W for sizes of interest to HTS devices. The reduction of cost from here tobelow $ 25/W is a very aggressive goal, and depends substantially on achieving an

Primary Goals Objectives OutcomesIncreased Efficiency (presentnominal 20%)

> 30% Carnot by 2005 Reduced operating expenses andmarket viability

Lower Capital Cost (presentnominal $100/watt)

< $25/cooling watt by 2007 withcryogenic components costing <10% total system

Reduced capital cost and marketviability

High reliability (present systemsdepend heavily on redundancy)

Operating availability > 99.8%by 2007

Mean time between failures ofoperating cryogenic system > 30years using redundancy andincreased component reliability

Secondary GoalsSize (present closed cycle systemsincluding auxiliary systems aremuch too large)

System & cryogenics 50%smaller by 2007

Utilization of full HTS systemsincreased power density

Variable cooling capacity (presentsystems might use stagedsmaller, less reliable, lessefficient, more expensiveequipment resulting in excess sizeand cost)

Cryogenics follow load usingstorage capacity or optimizedvariable speed drive techniques onHTS system by 2009

Significantly reduced penalty foroperating costs (utility cost ofbase load losses are 2.5-3 x largerthan losses which follow loadwith similar impact on industrialdemand charges)

Historic price decline withvolume and experience

Reduced costs as HTS systemspenetrate market

Commercial units at reasonablecost

Transparency (present MRIsystems suggest feasibility)

Customer acceptability by 2007 Low awareness of cryogenicsystem

Minimal Disruption (present MRIsystems suggest feasibility)

Customer acceptability by 2007 No interference of normaloperations

Maintainability (present MRIsystems suggest feasibility)

Customer acceptability by 2007 Average technicians can operatethe system

Soft failure mode (present MRIsystems suggest feasibility)

Customer acceptability by 2007 Cryosystem failures allowalternate operational schemes

Table 2. Cryogenic R&D needs expressed as goals, objectives, and desired outcomes.

06/18/01 9

economy of scale. Looking at the size range below 5 kW cooling capacity, withoutfundamental breakthroughs in cryogenics it is doubtful that even $ 50/W can be reacheduntil the market size exceeds 10,000 units/year.

As discussed by Mulholland et al8, there is a consistent relationship across themanufacture of many goods that shows how price falls with increasing volume. On log-log paper, as in figure 2, the exponent - 0.344 recurs again and again in the relationshipbetween price and quantity manufactured. That is,

$ (N1) / $ (N2) = (N1/N2) –0.344

Trusting in that relationship, it is plausible to expect price to drop from $ 100/W to$25/W if the number of cryogenic units manufactured increases by 2 - 3 orders ofmagnitude. That condition would correspond to widespread implementation of HTSdevices by utilities nationwide.

0.1

1

100 1000 10000 100000

Quantity per year

Per

un

it c

ost

rat

io

Figure 2 – Estimated relative costs of 60-80 K Cryocoolers vs quantity produced per year

Secondary Parametric Goals

There are additional performance parameters that are still very significant,although not as imperative as reliability, efficiency and cost. These include questions of:

06/18/01 10

Size – Can the cryogenics fit into a small space without requiring changes in the layout ofthe components it serves? How large is the refrigerator compared with the equipment itsupports? Is the combined HTS apparatus and cryogenic system smaller than aconventional system?Price History – Is the price declining steadily? Is it reasonable to believe about a certaincryogenic technology that it will be commercially viable within a few years?Transparency – Can the HTS device, including its cryogenics, be safely ignored (at leastfor very long periods of time)? Is this device a “turnkey” operation?Disruption – Does this device interfere with the normal way of doing things?Maintainability – Can utility employees of limited technical acumen easily carry out theneeded maintenance tasks, without extensive training?Failure Mode – Does this device fail in a “soft” way, giving adequate warning and/or notcausing a cascade of other problems? Is there an easy way to work around it?

The latter four fall into the broad category of “acceptability” to utilities.

Elements of the Transition

How will the transition from the present state to a better future state be made?As discussed in a later section, teams will likely carry out the R&D. However, it iscrucial to take note of the business conditions faced by any cryogenic manufacturer whotries to carry R&D results forward into commercial success.

There is a commonality associated with the introduction of new technology thatdeserves emphasis. Dr. Martin Nisenoff9 of Naval Research Laboratory has observed theway cost declines as military hardware evolves from R&D to large-volume production.Five stages can be identified:B. A. Unit purchase: This stage is simple: you buy whatever a vendor is selling. For a

200 W class refrigeration system, $ 104 is a representative figure. (For a kWcooling-power unit, $ 105 would be typical.)

C. Special Order: Some modification is involved at this stage. The purchaser goes toa vendor and specifies certain parameters, such as “a 20 Watt cooler” to operate ina certain range with specified efficiency, etc. For a price of the order of $105 ($30,000 - $300,000 ), the vendor delivers a device that meets the specifications,but isn’t really all that you wanted in the first place. It comes with no guaranteeand uncertain reliability.

D. R&D Project: At this stage, development is required. The purchaser’sspecifications are more precise and tailored to the application. The vendor ischosen carefully, based on prior performance. For example, on a spacecraft, youmight call for a cryocooler that runs for 5 years with no vibration; accordingly, itwill be bought from an aerospace company. The price range has risen by another

06/18/01 11

order of magnitude, perhaps between $300,000 and $3,000,000. This can befunded over a few years, because the units are being purchased one at a time.

E. Pilot Plant : At this stage, the price is up to about $107. The purchaser demandsexcellent reliability, so the vendor has to build a prototype manufacturing line toachieve this, and the output will be several cryocoolers. All the non-recurringengineering costs get paid here. But from this point on, the 20-watt cryocoolersonly cost $20,000.

It is important to understand that nobody is interested in a $106 market,because they can’t recover their investment. Also, nobody will commit $107 to apilot plant unless they are guaranteed a $107 market; an unproven market doesn’tattract investors. Who will build the first $107 prototype line? Without it, unitcosts won’t come down.

Another way of looking at this is to note that the “chicken and egg”problem is often a result of a market with a marginal business case. If theprojected profit margin is sufficiently large, businesses will enter. But if you haveto commit $107 to enter a market with 5% profit margins, it is very difficult toraise capital. Small manufacturers will even go after a $106 market, but only if theprofit margins are high.

Reliability doesn’t come until the vendor builds a lot of units. This is basicto the nature of manufacturing processes.

E. Full Scale Production: This is the desired end stage for production of cryogenicsystems. At this level, cryogenic systems for HTS devices in the national electricpower system would be routine. Today, for example, a typical major vendor isCTI in Massachusetts, the world’s biggest cryocooler manufacturer. Theirbusiness size is around $108 per year.

This sequence, drawn from military and aerospace experience, is historically valid formany types of technology. There is probably no way to get around this condition.

List of Issues

Any program of technology development must set its boundary conditions at theoutset. For the specific case of cryogenics to support HTS devices, there are an entireseries of unique considerations. One list, provided to the July 1999 workshop by NathanKelley and Jon Jipping, contains 20 issues that need to be addressed; they are somewhatcable-specific. The list appears here as Chart 1.

06/18/01 12

Chart 1 - Cryogenic System Issues for Utility Applications

1) HTS technologies are frequently sited for applications in congested regions. Therefore, the size of the refrigeration unit is acritical factor.

2) HTS cables’ long length to volume ratio and restricted diameter make the continuous flow of LN2 critical.

3) A refrigerator should be capable of operating for very long periods of time without external intervention or attention, as mostutility locations are unmanned.

4) System should be designed (and guaranteed) for 100% availability.

5) The load on a refrigerator will change depending on the electricity carried from a minimum (equal to the thermal inleak of thesystem) to a maximum design point. The system must operate efficiently through this entire range and reliably follow thecontinuously changing load.

6) Owing to their linear geometry, HTS cables have relatively high heat loads, and thus require an efficient refrigerator, not onlyat design load, but at all operating points. This is particularly significant for cables with daily and/or seasonal dips in loading.The overall system efficiency must be improved from the generally quoted 20W/W, as the life-cycle cost must be competitivewith conventional cables.

7) The system should be able to provide larger refrigeration capacity for a short duration following transient thermal conditions,such as short circuit.

8) A system needs to be “low profile”. This means that the refrigerator be as compact as possible. And, growing vertically is notalways a solution.

9) Most utilities would not permit a third party (i.e. a LN2 vendor) to make deliveries without their personnel being present. And,it is not feasible for a substation operator to be available every three days for the LN2 refill. Subsequently, evaporative bathcoolers are typically not an option as primary refrigerators.

10) Automatic circuit breakers and reconnects require some cycles to operate. The refrigerator cannot go through a completeshut-down/start-up cycle every time that the power “flickers”.

11) The power requirements for different installations could be very different. Despite the capacity differences, it is importantthat there be standardization of spare parts and repair techniques.

12) Other HTS technologies are operating below LN2 temperatures, and so will use He based systems. Different cycles orrefrigerator types for cables, transformers, etc. would require large parts inventory and diverse training.

13) Major components and long-lead items will need to be stocked, because a system cannot be out-of-service for several monthswaiting for replacement components.

14) The utility maintenance infrastructure will need training on routine and emergency maintenance.

15) For widespread commercial deployment of refrigeration systems, a skilled field-service force will be required.

16) System availability and reliability should be very high.

17) COST!!! Final system cost must be competitive on first installed and life-cycle basis.

18) Typical utility hardware has a 40-year depreciation. Refrigerator longevity will be compared to this experience.

19) The utilities have little experience with cryogenics and refrigeration. Therefore, they have many questions and doubts. Thecryogenics industry must be willing and able to work with system developers to educate the end users.

20) Remote control system capability should be integrated in the refrigeration system. Utilities are spending a lot of money toinstall remote monitoring and automation on their distribution systems. The refrigeration system must easily integrate with avariety of monitoring systems and protocols.

06/18/01 13

As mentioned above, this list is oriented toward transmission cables. Comparablelists (with considerable overlap) could be constructed for every other HTS application(generators, motors, transformers, fault-current limiters, etc.). The key point is that forany cryogenic system, there will be considerations of this type that have to be addressed.

Various Technology Paths

The available choices of cryogenic technologies can be broadly classed into threecategories: recuperative systems, regenerative systems, and hybrid systems. Thedifferent refrigerators have various advantages and disadvantages, which trade off againstone another in choosing the “best” cryogenic system for a particular application.

The reliability standards of utilities are notoriously stringent. Any cryogenicsystem that involves moving parts such as reciprocating pistons is immediately suspectedas unreliable. The attraction of some newer methods (such as pulse tube technology) isattributable in part to the freedom from moving parts in the cold end.

Recuperative Cycle Cryocoolers

The Turbo-Brayton cycle is the foremost unit of this type for reachingtemperatures of interest to HTS devices. One major advantage is that the transport fluidcan carry cold long distances, and this allows the cryogenics to be placed out of the wayin tight configurations. Another advantage is that the operating lifetime is long, because ituses gas bearings. Moreover, the flow is steady, and vibration is not a problem.

The disadvantages of the Turbo-Brayton cycle are that it requires a large heat-exchanger, and the unit is expensive to build. Most important, these systems cannot beminiaturized; the cost hits a plateau and doesn’t go any lower when the size declinesfurther. In the 77 K temperature range, that plateau comes at about 1000 watts ofrefrigeration power, too big for most HTS devices. Clearly, should R&D lead to smallerBrayton cycle refrigerators that retain good efficiency while lowering cost, it would be avery welcome improvement.

The Joule-Thomson cycle also bears mentioning. It too has steady, vibration-freeflow, and can transport cold fluids long distances. The absence of moving parts in thecold end is another advantage. However, it requires high pressure, and that typicallymeans oil-flooded compressors and the possibility of cold-head contamination.Improvements in clean compressor technology and in efficiency (in the 60-80K range)would make these systems more attractive for HTS cooling.

Joule-Thomson systems using mixtures of gases (N2, H2, He) are used to reachtemperatures well below 77 K. Mixed-gas JT refrigerators are able to achievetemperatures around 77 K using conventional air conditioning compressors, which hasbeen a major advantage of the mixed gases. They also give much higher efficiency. Thehigh pressures are only required for use with pure nitrogen, which has a low efficiency for

06/18/01 14

reaching 77 K. Current technology is limited to > 90K if competitive efficiency is to beobtained.

Regenerative Cycle Cryocoolers

There are three devices1 (all cousins of one another) that are the leading candidatesin this category: Stirling Cycle, Gifford McMahon, and Pulse Tube refrigerators. All threework by having a transport fluid (a gas) pass cyclically through a regenerator and adisplacer. The displacer moves back and forth at low temperatures. These systemsoperate at frequencies below 60 Hz at 77 K, and as slow as 1 Hz when cooling to belowthe 4 K range. The advantage of the Pulse Tube device is that the “displacer” is made out of acolumn of gas, not solid material – it is a gas plug. This eliminates a crucial moving part atlow temperatures, and greatly enhances reliability. Most pulse tube cryocoolers built todate have had small cooling capacities (50 W or less), but recent advances havedemonstrated the feasibility of systems with up to 1 kW of cooling capacity at 77 K, andmuch larger capacities are not unreasonable in the future. Pulse tube cold heads have alsobeen used with thermoacoustic engines to build natural-gas-fired industrial gas liquefiers.Such systems offer the possibility of high reliability due to the lack of moving parts ineither the driver or the cold head, however the current technology results in a physicallylarge cryocooler with limited efficiency. Rapid advances are being made in the pulse tubeand thermoacoustic cooler fields, and the technology shows long-term promise.

The Stirling Cycle has several advantages, notably high efficiency, small size andweight, and moderate cost. There is considerable manufacturing experience with suchunits; over 100,000 have been made already. On the other hand, there is always going tobe vibration, owing to the moving displacer. It must be run dry (without lubrication)because of the links (via the coolant gas) between cold and warm regions. Moreover, it isexpensive to achieve long lifetimes (3 – 10 years) in these systems.

The Gifford-McMahon cryocooler is the most popular of this type. On the onehand, it isolates the compressor from the regenerator and displacer, which allows amodified air-conditioning compressor to be used. This keeps the cost down. About20,000 units are made per year, so there is plenty of experience, and reliability (assumingregular maintenance) is good. Maintenance is required about every one to three years.Lifetime of compressor and valve parts may be about 5 years. Reliability may be good,but not good enough for the utility industry. On the other hand, the efficiency is muchlower than in the Stirling cycle, expressly because an external AC compressor is used.There is still inherent vibration from the moving displacer. A Gifford-McMahon unit islarge and heavy, but this problem is mitigated somewhat because the compressor can beplaced some distance away from the place where cooling must occur.

Hybrid Open Loop Systems

06/18/01 15

Given that most electric-power applications are expected to operate near 77 K, itis useful to consider systems based on liquid nitrogen, coupled with an auxiliary device tocool slightly further.

The foremost advantage of LN2 is that it is so cheap. As a byproduct of oxygen-production plants, the available supply of LN2 so outweighs the demand for it that theprice is very low – about six cents per liter in truckload quantities. This implies thatsomeone else has paid the thermodynamic and economic cost of reaching 77 K. Thecapital cost of basing a system on LN2 must include the cost of a storage tank; theoperating cost includes the cost of trucking in LN2 occasionally. The severity of thereliability criterion sets the size (and cost) of the storage tank. These additional costburdens are small compared to the cost of electric power to reach 77 K in any other way.Wherever other criteria (geometry, location, etc.) do not mandate otherwise, liquidnitrogen is the coolant of choice.

To get to some lower temperature, several avenues are possible:• First, pumping a vacuum on LN2 can reduce its temperature from 77 K down to

64 K. Of course, the reliability of the vacuum pump needs to be considered; itmay not be as good as most cryocoolers.

• Second, a mechanical cryocooler running from an “ambient” platform of 77 K doesnot have to work very hard, compared to one running from room temperature.

• Third, helium gas can be used with another refrigeration system to reach muchlower temperatures.

Any of these combinations can properly be termed a “hybrid” system. One variation would be to have a cryocooler maintain a bath of LN2 at 77 K. Thatwould provide redundancy, and average the load over some time period. If the coolerfailed, there would then be a back-up reservoir that would keep things cold for some time,relying on the transport of LN2. Trade-offs play a large role in optimizing a hybrid system. For example, supposeit is desirable to have a HTS device operate at 70 K, but it would still be able to functionat 77 K, although limited in some way. In the event of 77 K operation, performancewould be lessened, but reliability would not be lost. It is plausible that a utility mightfind that acceptable, and the system designer could save the cost of redundancy. Issuesof this type need to be considered when designing a hybrid system. Candidate Technologies The question “which class of cryogenic refrigerator is most likely to be used witheach application?” needs to be addressed; obviously the actual choice lies with the teampursuing the development of a cooler. However, it is relatively easy to identify certaintechnologies that initially appear more appropriate to one or another HTS device. (Forexample, regarding transmission lines, one may ask whether to employ a refrigerator orsimply use liquid nitrogen from a dewar.) Table 3 below presents some suggestedrefrigeration cycles for each of the leading HTS applications. To determine the “best”technology for each case, it is useful to examine obstacles from the point of view of a

06/18/01 16

problem statement and a proposed action. Appendix D offers one type of form10 thatcan be used to facilitate the identification and resolution of a problem.

06/18/01 17

Table 3: Candidate refrigeration cycles for HTS applications

HTS Application Proposed cycle Alternate Cycle Transmission Lines Reversed-Brayton Hybrid Transformers Gifford-McMahon Pulse Tube Motors Reversed Brayton Hybrid Fault current limiters Gifford-McMahon Pulse Tube Superconducting Magnetic EnergyStorage

Gifford-McMahon Pulse Tube

Flywheel Energy Storage Gifford-McMahon Pulse Tube Magnetic Separators Gifford-McMahon Pulse Tube

Along the various technical pathways, some of the questions that will be asked ofeach candidate technology are these:ß Which components of the cryogenic refrigeration system are behind in development

for achieving capacity targets?ß Which components of the cryogenic refrigeration system are behind in development

for achieving efficiency goals?ß What is the reliability for existing components of the cryogenic refrigeration system?

Moreover, it will be important to develop realistic refrigerator models, using data fromreal-world components, in order to design systems that achieve the desired capacities.

Research Avenues

The Superconducting Partnership Initiative (SPI) approach has proved verysuccessful in bringing HTS out of the laboratory and applying it in actual devices for theelectric power industry. It is plausible to think that this method will likewise work wellfor cryogenic systems associated with these devices. The strongest feature of SPIprojects is that they are industry-driven. Here the government sponsor pays only halfthe cost (industry must raise the rest), and industry keeps the patent protection on newinventions. As a result, it is guaranteed that industry remains interested in and attentiveto the progress of the work throughout the entire duration of the project. This stands insharp contrast to an older way of doing R&D, in which government does it all on thefront end, and then industry gets into the act only much later – by which time the largegovernment effort may no longer be particularly applicable to the commercialmarketplace.

Therefore, this Roadmap strongly recommends that development of cryogenicsystems for HTS devices be conducted via programming analogous to theSuperconducting Partnership Initiative.

06/18/01 18

Roles of the Participants

In a true collaboration, each partner participates at all stages, although one oranother may assume dominant responsibilities in different phases. To assemble a winningteam, it is necessary for all the different specialists to work together as a team, dedicatedto reaching the goal together. Just as for the HTS devices now being constructed withinthe SPI program, similarly any successful venture toward better cryogenic systems mustcontain a collection of very capable individual players:F. G. Industry

The economic history of America shows that things don’t get done unlesssomebody makes a profit. At the outset, this roadmap defines the goal as “commercialsuccess”, and recognizes that profitability drives business decisions. If there is ever to becommercial success for this program, it will be because industry leads the effort. Thefocus on business goals – profitability – leads to keeping costs down, to products that arecapable of being manufactured in quantity (not just laboratory apparatus), and toperformance that meets the customers’ needs.

When the stockholders are asked to lay their money on the line, there must be animplicit promise that their business goals will be respected, and their decision-makingprocesses given deference in planning an R&D strategy. This is why no realistic roadmapcan specify in detail the path to accomplishments. People who have been in thecryogenics business have a strong sense of what works and what doesn’t; their leadershipis essential to success here.

There is already a very substantial cadre of cryogenic system manufacturers inplace nationally. The new requirements of HTS devices to be deployed in utility systemschallenges these companies to build systems that are superior in many ways to existingequipment. The companies are also being asked to take a risk that a major market willdevelop over time for their new products, so that their initial development cost can berecovered. People willing to commit their own capital and take such risks deserve to bethe ones who call the shots. Ultimately, industry will choose the manufacturing goals,define a practical timetable, and establish the go/no-go decision points.

H. Utilities

Only in the last few years have utilities begun to participate in energy R&D,through CRADAs and SPI programs. By providing a test-bed where hardware can beinstalled in real electric power systems, they have very quickly converted researchscientists into believers in the very special circumstances under which new devices mustfunction. It is eminently clear that their voice must be listened to carefully, because theyare the ones who ultimately must decide whether to commit to installation of HTSdevices. There is a go/no-go decision point in the loop that researchers often overlook.

06/18/01 19

Throughout its 100-year history, the electric utility industry has faced numeroustechnical and societal challenges. Today, changes in federal and state regulation promise tofundamentally alter the relationship between utilities and their customers. As such, everymajor utility in the US is engaged in activities to improve productivity and efficiency,thereby lowering costs. At the same time, with rapid advances in and proliferation ofpersonal computers and a host of electronic devices, highly reliable electric service is thesingle most important feature for customers. Recent events during the summer of 1999 inChicago and New York illustrate the "zero tolerance" attitude many have toward poorreliability. Utility managers will not commit to install unproven technology on the electricsystem. This is especially true from a reliability standpoint, but also from a costperspective.

The foremost role of the utility in an R&D partnership is to keep the focus of allparticipants on the ultimate goal. Even at the initial design stage, scientists mustrecognize that their conceptual freedom is limited by very practical requirements. Thereis no point in starting down a path toward a device that will eventually demand constantattention. By heeding the advice of the utility engineers at early stages, the probability ofacceptance later on is greatly enhanced.

When the device is ready to be “rolled out”, it is the utility staff, working with the industrialpartner, who will test and refine the entire unit within its operating environment. Too often in past R&Dprograms, this has been treated as an afterthought, but such is not going to be the case in HTS technology.The entire system, including the cryogenic system, has to function together, reliably. It is the utilitypartner who will dominate the decision about how that goal is to be achieved

I. National Laboratories

The highest scientific capabilities in the HTS program reside within the NationalLaboratories. The Labs have proven over half a century that they have the ability toaccomplish very difficult scientific tasks, identifying fundamental obstacles andovercoming them in a scientifically sound manner. In recent years, the labs havedemonstrated their ability to work collaboratively with industry through CRADAs. Inthis roadmap, we envision that the labs will contribute across the board, but especially infundamental areas such as materials science, properties of gases, and other technologieswhere industry is unlikely to take a leading role.

For the special case of cryogenics to support HTS devices within utility systems,the present list of obstacles is formidable: we seek an increase in efficiency of 50% abovetoday’s level, and we seek to elevate mechanical equipment to a level of reliabilityheretofore seen only in entirely passive, non-moving components. Even to place thesetopics on the agenda is to indicate how much confidence we have in the cleverness andresourcefulness of national lab scientists and engineers.

J. Other Government Agencies

06/18/01 20

The National Institute of Standards and Technology (NIST), as well as NASA andthe Department of Defense (especially the Naval Research Laboratory) have had longexperience with cryogenic technology. DoD and NASA have taken laboratory devices allthe way to reliable satellite systems, and NIST is the leader in most basic areas ofrefrigeration research1,11. It is absolutely essential to the success of this effort that theseagencies remain strongly coupled into the program – both because of their expertise inresearch, and their appreciation of what is required of a cryogenic system operating in thereal world.

The very mission-specific nature of most NASA and DoD programs means thatthey will be complementary to, not overlapping, this work relating to HTS devices forelectric utility systems.

Timetable

At this point in time, it is impossible to be either prescriptive or precise in layingout a timetable for cryogenic R&D in support of HTS devices. However, certain generalfeatures can be discussed, and the broad scope of the activity delineated:

Fiscal year 2000 is essentially finished. This has been a time of organizing andplanning, and this roadmap is part of that process. During this year, continuingdiscussions among likely participants has to a consensus on the nature of the R&D to bepursued but has also opened other areas for consideration.

In 2001 (fiscal year or calendar year), teams will be assembled, agreements will bereached, and participants will get establish their financial mechanisms. Given the typicaltime-cycle of government authorizations and appropriations, it would not be credible tothink that funding would be available any sooner than the start of FY2002. Anincremental appropriation of $10 million as the government’s share is about the right sizeto get several analogous SPI projects (Cryogenic Partnerships for Superconductivity)started.

Accordingly, actual work would start in 2002. As discussed above, the utilitypartners would be only peripherally active in the early stages, while industry and nationallab researchers would play a leading role.

This phase of exploring new avenues and trying new technological innovationswould last through 2003. The industrial collaborator would certainly be the dominantpartner by late 2003. Bench testing and prototype refinements would be the mainactivity. Some go/no go decision points during 2003 will result in the discontinuation ofsome projects and greater emphasis on others.

In 2004 and 2005, implementing real systems would be the principle theme of theprogram. Making something work in a utility environment will call for the fullcooperation of all partners, from bench scientists to utility engineers, refining the design,the manufacturing process, and the way it is incorporated into the overall system. Thegovernment role would be greatly diminished by the conclusion of 2005, and the utilities’role greatly enhanced. Here again, business decisions will be made that further narrow thechoices among candidate technologies.

06/18/01 21

2006 should be a time of final refinements, leading to proven hardware that meetsthe original design goals of reliability, efficiency and cost. The entire program will bealmost entirely in the private sector by this time, with national lab scientists playing onlyan advisory role.

It is an interesting coincidence that in a recent study of the economics of futureHTS applications by Mulholland et al12, the estimated year of first market-entry formany of the HTS devices is 2007.

Obviously, if some of the existing SPI projects prove highly successful in the nextcouple of years, there will be an incentive to accelerate a cryogenics R&D program. It isplausible to think that an "SPI" for refrigerators could follow upon the present SPI IIprojects. Imagine a specific example: after the 12-18 month operation of the SPI II HTSCable demonstration at Detroit Edison, an "SPI III" refrigerator project could install a newand advanced refrigerator at the same site. Much will be learned from the HTS Cableproject about the refrigeration system. Subsequently, any SPI II project could be used asa "retrofit" demonstration for an improved refrigeration system.

Summary and Conclusions

This roadmap is offered to the U.S. Department of Energy by the assembly ofutility, industry and national lab researchers who believe that now is the right time todevelop cryogenic systems that will allow HTS devices to become standard componentsof electric power systems. Utility equipment is the sum of its parts; in this case, bothsuperconductors and cryogenics are required to make the system work. Advances inboth the HTS materials and in refrigeration technology (either applied or new) willdetermine the eventual level of success of these products.

There is widespread agreement that much science is yet to be learned insuperconductivity. On the other hand cryogenics has been dismissed as working out afew engineering details to meet the specific needs of the application, and that can behandled mostly by industry. This roadmap opposes that line of thinking. We feel ittakes much more than simply providing a few engineering details. To get from where weare now to where we need to be, there is need for much new science in cryogenics toimprove the efficiency and reliability and to reduce costs. Understanding how to providethe required refrigeration with fewer and more reliable moving parts and having theprocess work efficiently requires research.

Because of two previous workshops, this document devotes only a little space toreviewing where we are today. Much more important is where we want to be – where wemust get to if commercial success is to be achieved. We establish specific numericalperformance goals for cost, efficiency, and reliability. We examine the various possiblepathways to reach these performance goals, taking note of both the technologies that canbe pursued, as well as the business constraints that must be considered. We alsorecommend an R&D strategy of forming teams of knowledgeable specialists, in order toimprove the probability of achieving success. Finally, we suggest a typical timetablespanning several years over which these activities would take place.

06/18/01 22

We expect the benefits of this program to be very widespread. Concepts thatincrease the efficiency, improve reliability, and reduce costs would apply to almost allareas of superconducting power applications, as contrasted to only individual devicesdeveloped by SPI projects. This cryogenics R&D would be an “umbrella” program inparallel with other ongoing efforts.

It is too early to specify a detailed research plan, and trying to “pick winners”has always proved futile. Rather, our purpose in this roadmap is to identify the broadscope of collaborative R&D effort involving partners drawn from industry, utilities andthe national labs. The details are to be filled in by those (in the private sector) whobelieve in their potential accomplishments enough to risk their own money in thisendeavor. The program described here is definitely an industry-led enterprise.

References 1 R. Radebaugh, Refrigeration for Superconductors, presented at Applied

Superconductivity Conference, September 19982 R.F. Barron, Cryogenic Systems (pp. 60-82) (Oxford Univ. Press: 1985)3 T. P. Sheahen, Introduction to High Temperature Superconductivity (Plenum

Press: 1994)4 J. W. Mulholland, et al., Analysis of Future Prices and Markets for High

Temperature Superconductors, Appendix 6 (U.S. Dept. of Energy publication,2000)

5 R.D. Blaugher, IEEE Spectrum, (July 1997), pp. 36-426 F.J. Kadi and R.C. Longsworth, ASME Journal of Engineering for Industry

(August 1977)7 J.L. Martin, Mesoscopic Devices, priv. comm.8 J. W. Mulholland, et al., Analysis of Future Prices and Markets for High

Temperature Superconductors, Appendix 15 (U.S. Dept. of Energy publication,2000)

9 M. Nisenoff, NRL, priv. comm. 10 J. A. Demko, ORNL, priv. comm.11 See, for example, R. Radebaugh “Development of the Pulse Tube Refrigerator as

an Efficient and Reliable Cryocooler”, Proc. Institute of Refrigeration (London),1999-2000

12 J. W. Mulholland, et al., Analysis of Future Prices and Markets for HighTemperature Superconductors, (U.S. Dept. of Energy publication, 2000).